CN111598771B - PCB (printed Circuit Board) defect detection system and method based on CCD (Charge coupled device) camera - Google Patents

Abstract

The invention provides a PCB circuit board defect detection system based on a CCD camera, including a CCD image acquisition module, and an image recombination splicing module, a defect detection module, and a self-calibration module respectively arranged on the same host computer as the CCD image acquisition module Module, based on the above system, the present invention also discloses a PCB circuit board defect detection method based on a CCD camera. The present invention utilizes the feature of line-by-line scanning of linear CCDs to collect images, and realizes partial image information detection and information adjustment by controlling the number of image imaging lines. The invention uses a self-correction module to adjust the number of spliced lines to obtain image line number information and avoid false detection. The invention effectively utilizes the linear CCD camera to effectively improve the efficiency of PCB defect detection, save production cost and improve real-time performance.

Description

A PCB circuit board defect detection system and method based on CCD camera

technical field

The invention belongs to the technical field of PCB circuit board defect detection, and in particular relates to a CCD camera-based PCB circuit board defect detection system and method.

Background technique

There are many processes in the PCB manufacturing process, and defects may occur in each process. If these subtle defects cannot be found accurately and quickly in the production process, it will cause a decline in the product qualification rate and affect its reliability. It may even The entire printed board is scrapped, which increases the production cost. If a fault is found on the finished PCB circuit board, the cost is huge, and the cost of putting the faulty PCB board on the market is even more fatal. Therefore, defect detection plays a very important position in the PCB circuit board production process. At the same time, in the existing on-line detection method of PCB circuit board, there is a bottleneck in the detection speed. Therefore, the present invention utilizes the characteristics of scanning and imaging of the linear CCD camera to greatly improve the detection efficiency and save the cost at the same time.

Contents of the invention

Aiming at the above-mentioned deficiencies in the prior art, the present invention provides a PCB circuit board defect detection system and method based on a CCD camera, which can improve real-time acquisition and detection efficiency, and has high accuracy and real-time performance, so as to realize PCB circuit board defect detection system and method. Further development of board in-line inspection.

In order to achieve the above object, the technical scheme adopted in the present invention is:

This solution provides a PCB circuit board defect detection system based on a CCD camera, including a CCD image acquisition module, and an image reorganization splicing module, a defect detection module and a self-correction module that are respectively set on the same host computer as the CCD image acquisition module ;

The CCD image acquisition module is used for real-time scanning and acquisition of images, and the images collected by real-time scanning and the standard sampling images obtained are transferred to the image reorganization stitching module; the CCD image acquisition module includes a CCD camera and a CCD image acquisition interface;

The image recombination and stitching module is used to receive the image line number information imported from the correction module, and according to the image line number information, perform image splicing and reorganization on the input real-time collected images and standard sampling images by line, and splicing The final image is passed to the defect detection module;

The defect detection module is used to calculate the brightness, contrast and structure of the corresponding image according to the spliced collected image and the standard sampled image using a structural similarity algorithm, and obtain the difference between the collected image and the standard sampled image according to the calculation result, and according to The difference obtains the defect position, and transfers the defect position to the self-calibration module;

The self-correcting module is used to adjust the number of scanning lines according to the obtained defect position, and input the adjusted number of lines to the image stitching module.

Based on the above method, the present invention also discloses a method for detecting defects of a PCB circuit board based on a CCD camera, comprising the following steps:

S1. Obtain a standard sampling image, and set the number of collection lines of the CCD camera and the basic settings of the CCD camera, and use the CCD camera to scan and collect images in real time;

S2. Input image row number information, and according to the image row number information, perform image splicing and reorganization processing on a row-by-row basis for the standard sampled image and the real-time collected image;

S3. Using the structural similarity algorithm to calculate the brightness, contrast and structure of the spliced collected image and the standard sampled image respectively, and judge whether there is a defect according to the calculation result, if so, enter step S4, otherwise, end this detection, thereby Complete the defect detection of PCB circuit board;

S4. Perform the processing of adjusting the number of scanning lines according to the judgment result, and use the adjusted information on the number of scanning lines as the information on the number of lines of the input image in step S2, and return to step S2.

Further, the matrix expression of the real-time image acquisition in the step S1 is:

The matrix expression of the standard sampling image is:

Among them, A is the matrix of real-time collected images, B is the matrix of standard sampled images, ε is the starting scanning position, x is the number of rows per scan, y is the number of columns per scan, and η is the standard sampled image intercepted starting value.

Still further, the step S2 includes the following steps:

S201, input real-time acquisition image;

S202. Input image row number information: set a default value of the number of stitched rows, and calculate the value of the next stitched row according to the default value of the number of rows;

S203. Image splicing and reorganization: according to the default value of the number of spliced rows, perform image splicing and reorganization processing on a row-by-row basis for the standard sampled image and the real-time captured image.

Still further, the expression of the value of the next spliced row in the step S202 is as follows:

ω′＝ω+R _pre +R _after

Among them, ω' is the value of the next spliced row, R _pre is the required previous row, R _after is the required subsequent row, and ω is the default value of the spliced row number.

Still further, the matrix expression of the recombined real-time acquired image in the step S203 is:

The matrix expression of the standard sampling image after the recombination is:

Among them, A' is the reorganized real-time acquisition image matrix, B' is the reorganized standard sampling image matrix, R _pre is the required previous row, R _after is the required subsequent row, ω is the default value of the number of spliced rows, and x is The number of rows for each scan, y is the number of columns for each scan, and η is the initial value of standard sampling image interception.

Still further, the step S3 includes the following steps:

S301. Calculate and obtain the brightness, contrast and structure of the spliced real-time captured image and the standard sampled image respectively by using a structural similarity algorithm;

S302. Proportionally fuse the brightness, contrast and structure of the real-time collected image and the standard sampled image to obtain an evaluation function;

S303. Judging whether the evaluation function is greater than the preset detection threshold T _d , if so, end this detection, thereby completing the defect detection of the PCB circuit board, otherwise, mark the defect position in the currently collected image, and output the current Defect detection image, and go to step S4.

Further, the expression of the evaluation function is as follows:

F(A',B')=[L(A',B')] ^α [C(A',B')] ^β [S(A',B')] ^γ

为矩阵A'的像素平均灰度值的平方，

为矩阵B'的像素平均灰度值的平方，C₁，C₂，C₃均为用来增加计算结果的稳定性参数，C(A',B')为矩阵A'和矩阵B'的对比度对比函数，

为矩阵A'的方差，

为矩阵B'的方差，S(A',B')为矩阵A'和矩阵B'的结构对比函数，σ_A'B'为矩阵A'和矩阵B'的协方差，α，β，γ均为调整三个模块间的参数，A'为重组后的实时采集图像矩阵，B'为重组后的标准采样图矩阵。Among them, F(A', B') is the evaluation function, μ _A' is the pixel average gray value of matrix A', μ _B' is the pixel average gray value of matrix B', N is the total number of pixels, x _i is the value of the pixel point corresponding to matrix A', y _i is the value of the pixel point corresponding to matrix B', i is the subscript of the corresponding point in matrix A', σ _A' is the standard deviation of matrix A', σ _B' is the standard deviation of matrix B', L(A',B') is the brightness comparison function of matrix A' and matrix B',

is the square of the average gray value of the pixel of the matrix A',

is the square of the average pixel gray value of matrix B', C ₁ , C ₂ , and C ₃ are all stability parameters used to increase the calculation results, and C(A', B') is the matrix A' and matrix B' contrast contrast function,

is the variance of the matrix A',

is the variance of matrix B', S(A', B') is the structure comparison function of matrix A' and matrix B', σ _A'B' is the covariance of matrix A' and matrix B', α, β, γ Both are to adjust the parameters among the three modules, A' is the real-time acquisition image matrix after reorganization, and B' is the standard sampling image matrix after reorganization.

Still further, the step S4 includes the following steps:

S401. Judging whether there is a defect detection image input according to the judgment result, if yes, then enter step S402, otherwise, end the process;

S402. Determine whether the previous scanning line information is needed, if so, set the current line number information as the required previous line number R _pre , and enter step S403, otherwise, the imaging range of the current image is the initial range of the PCB circuit board, And enter step S403;

S403, judge whether to need the scanning row information afterwards, if so, then set the current row number information as the required row number R _after , and use the required row number R _after to complete the current image information, and enter step S404, otherwise , end the process;

S404. Use the required number of rows before R _pre and the number of rows _{after R after} as the input image row number information in step S2, and return to step S2.

Furthermore, the expression of the line number information of the input image in the step S404 is as follows:

ω′＝ω+R _pre +R _after

Among them, ω' is the line number information of the output image, that is, the value of the next spliced line, R _pre is the required previous line, R _after is the required subsequent line, and ω is the default value of the spliced line number.

Beneficial effects of the present invention:

(1) In order to overcome the deficiencies in the real-time and efficiency of the existing PCB online detection technology, the present invention uses a CCD as a collection device, uses a PCB circuit board as a detection object, and supports technologies such as image collection and image processing to carry out PCB board inspection. Online defect detection can improve the efficiency of real-time collection and detection, and has high accuracy and real-time performance, so as to realize the further development of PCB online detection. The present invention is carried out on the basis of using a linear CCD camera at a standard acquisition speed, utilizing the characteristics of the linear CCD camera to scan and acquire images by row, and using the structural similarity algorithm as the basis for defect detection to detect defects on PCB circuit boards , because the use of structural similarity algorithm for defect detection requires a large amount of image information, and the commonly used one is based on the overall image. The present invention utilizes the characteristics of a CCD camera to realize partial image information detection and information adjustment by controlling the number of image imaging lines;

(2) The present invention uses a linear CCD to collect images, and utilizes the characteristics of linear CCD line scanning imaging to realize image mosaic, defect detection, and self-correction functions. At the same time, the CCD acquisition module and other modules do not interfere with each other and run in parallel to realize online detection. Effect. In the present invention, the collected image and the standard sampled image are respectively transmitted to the image restoration module, the defect detection module, and the self-correction module to complete the defect detection of the PCB, and the self-correction module is used to adjust the number of stitching lines, complete the image information, and avoid false detection . The invention effectively utilizes the linear CCD camera to effectively improve the efficiency of PCB defect detection, saves the production cost and improves the real-time performance.

Description of drawings

Fig. 1 is a system block diagram of the present invention.

Fig. 2 is a parallel schematic diagram of an embodiment of the present invention.

Fig. 3 is a flow chart of the method of the present invention.

Fig. 4 is a flow chart of the image reorganization and stitching module of the present invention.

Fig. 5 is a flowchart of the defect detection module of the present invention.

Fig. 6 is a flowchart of the self-calibration module of the present invention.

Detailed ways

The specific embodiments of the present invention are described below so that those skilled in the art can understand the present invention, but it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art, as long as various changes Within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are included in the protection list.

Example

As shown in Fig. 1, the present invention provides a kind of PCB circuit board defect detection system based on CCD camera, comprises CCD image acquisition module, and the image reorganization splicing module that is respectively arranged on the same upper computer with described CCD image acquisition module, Defect detection module and self-correction module;

The CCD image acquisition module is used for real-time scanning and acquisition of images, and the images collected by real-time scanning and the standard sampling images obtained are transferred to the image reorganization stitching module; the CCD image acquisition module includes a CCD camera and a CCD image acquisition interface;

The image recombination and stitching module is used to receive the image line number information imported from the correction module, and according to the image line number information, perform image splicing and reorganization on the input real-time collected images and standard sampling images by line, and splicing The final image is passed to the defect detection module;

The defect detection module is used to calculate the brightness, contrast and structure of the corresponding image according to the spliced collected image and the standard sampled image using a structural similarity algorithm, and obtain the difference between the collected image and the standard sampled image according to the calculation result, and according to the obtained The defect position is obtained from the above difference, and the defect position is passed to the self-calibration module;

The self-correcting module is used to adjust the number of scanning lines according to the obtained defect position, and input the adjusted number of lines to the image stitching module.

In this embodiment, the present invention adopts a parallel module, and the CCD image acquisition module works in parallel with the image reorganization and splicing module, the defect detection module and the self-correction module to improve the real-time performance of detection and complete online detection.

In this embodiment, as shown in FIG. 2 , the operation of the whole system mainly includes two parts, that is, the collection program and the detection program. Both the collection program and the detection program are designed to run in parallel on the same host computer. The acquisition program mainly runs the image acquisition module, and the detection program mainly runs the image splicing and recombination module, the defect detection module and the self-correction module. From the perspective of the time axis, the acquisition program and the running program run in parallel, because the linear CCD acquires images through line scanning, so in the process of line scanning, the detection program mainly splices and detects certain lines to form an independent program block . Each detection program block runs independently without interfering with each other, and runs in a parallel manner. From the perspective of the axis of image acquisition lines, each detection program detects different lines to achieve parallel operation.

The present invention is carried out on the basis of using a linear CCD camera at a standard acquisition speed, utilizes the characteristics of the linear CCD to scan and acquire images by lines, and uses the structural similarity algorithm as the basis for defect detection to detect defects on PCB circuit boards. Since the use of structural similarity algorithm for defect detection requires a large amount of image information, the commonly used ones are mainly the overall image. The present invention utilizes the characteristics of linear CCD to scan and collect images by lines, and realizes partial image information detection and detection by controlling the number of image imaging lines. The information is adjustable, as shown in Figure 3, and its implementation method is as follows:

S1. Obtain a standard sampling image, and set the number of acquisition lines of the CCD camera and the basic settings of the CCD camera, and use the CCD camera to scan and acquire images in real time.

In this embodiment, according to the characteristics of linear CCD imaging, the number of rows of each scan is set to x, the number of columns is y, the number of columns is set to a fixed value, and the initial scanning position is set to ε, so the image scanned each time is The size is x×y, and the resulting matrix is:

At the same time, the input size of the required standard sampling image is also x×y, and the input size of the standard sampling image is the intercepted part of the number of pixel rows in the image itself, and an original standard sampling image is retained to intercept the starting value η, so the image is intercepted The imaging extent matrix of :

Pass the real-time acquisition image matrix A to the image reconstruction and stitching module.

S2. Input image row number information, and according to the image row number information, perform image splicing and reorganization processing on the standard sampling image and real-time collected image by row, as shown in Figure 4, the implementation method is as follows:

S201, input real-time acquisition image;

S202. Input image line number information: set a default value of the number of spliced lines, and calculate the value of the next spliced line according to the default value of the number of spliced lines;

The expression for the next concatenation row value is as follows:

ω′＝ω+R _pre +R _after

Among them, ω' is the value of the next spliced row, R _pre is the required previous row, R _after is the required subsequent row, and ω is the default value of the spliced row number;

S203. Image splicing and reorganization: according to the default value of the number of spliced rows, perform image splicing and reorganization processing on the standard sampled image and the collected image by row, and the matrix expression of the real-time collected image after reorganization is:

The matrix expression of the standard sampling image after the recombination is:

Among them, A' is the reorganized real-time acquisition image matrix, B' is the reorganized standard sampling image matrix, R _pre is the required previous row, R _after is the required subsequent row, ω is the default value of the number of spliced rows, and x is The number of rows for each scan, y is the number of columns for each scan, and η is the initial value of standard sampling image interception.

In this embodiment, the image reorganization and stitching module is designed to stitch the input real-time captured images line by line and transmit them to the defect detection module. At the same time, the image reorganization and stitching module also receives the information on the number of stitching lines delivered by the self-calibration module, which is convenient for the next step. Detection processing.

S3. Using the structural similarity algorithm to calculate the brightness, contrast and structure of the spliced real-time acquisition image and the standard sampling image respectively, and judge whether there is a defect according to the calculation result, if so, enter step S4, otherwise, end this detection, In order to complete the defect detection of the PCB circuit board, the implementation method is shown in Figure 5:

S301. Using the structure similarity algorithm to calculate the brightness, contrast and structure of the spliced real-time collected image and the standard sampled image respectively;

S302. Fusing the brightness, contrast and structure of the real-time collected image and the standard sampled image in proportion to obtain an evaluation function;

S303. Judging whether the evaluation function is greater than the preset detection threshold T _d , if so, end this detection, thereby completing the defect detection of the PCB circuit board, otherwise, mark the defect position in the currently collected image, and output the current Defect detection image, and go to step S4.

In this embodiment, the defect detection module is designed to calculate the corresponding brightness, contrast and structure of the incoming collected image and the standard sampling image through the structural similarity algorithm, and obtain the difference between the two images through specific calculations. , so as to obtain its defect position, if there is a defect, it will be marked and passed to the self-calibration module.

In this embodiment, from the image reorganization and stitching module, the incoming acquisition image matrix A' and the incoming standard sampling image matrix are B'. Image quality is restricted by brightness information and contrast information. Therefore, when calculating image quality, it is necessary to consider the influence of the two while considering structural information.

In this embodiment, the average gray value of the image is used as the estimation of the brightness measurement, and the average gray value of the pixel of the matrix A' is:

The pixel average gray value of matrix B' is:

Taking the standard deviation of the image as the contrast estimation value, the standard deviation of matrix A' is:

The standard deviation of matrix B' is:

The brightness contrast function of matrices A' and B' is:

The contrast function of matrices A' and B' is:

The structure comparison function of matrices A' and B' is:

组合亮度，对比度和结构三个相关函数，得到评价函数：in,

Combining the three correlation functions of brightness, contrast and structure, the evaluation function is obtained:

F(A',B')=[L(A',B')] ^α [C(A',B')] ^β [S(A',B')] ^γ

为矩阵A'的像素平均灰度值的平方，

为矩阵B'的像素平均灰度值的平方，C₁，C₂，C₃均为用来增加计算结果的稳定性参数，C(A',B')为矩阵A'和矩阵B'的对比度对比函数，

为矩阵A'的方差，

为矩阵B'的方差，S(A',B')为矩阵A'和矩阵B'的结构对比函数，σ_A'B'为矩阵A'和矩阵B'的协方差，α，β，γ均为调整三个模块间的参数，A'为重组后的实时采集图像矩阵，B'为重组后的标准采样图矩阵。In the above formula, F(A', B') is the evaluation function, μ _A' is the pixel average gray value of matrix A', μ _B' is the pixel average gray value of matrix B', N is the total number of pixels, x _i is the value of the pixel point corresponding to the matrix A', y _i is the value of the pixel point corresponding to the matrix B', i is the subscript of the corresponding point in the matrix A', σ _A' is the standard deviation of the matrix A', σ _B' is the standard deviation of matrix B', L(A', B') is the brightness contrast function of matrix A' and matrix B',

is the square of the average gray value of the pixel of the matrix A',

is the square of the average pixel gray value of matrix B', C ₁ , C ₂ , and C ₃ are all parameters used to increase the stability of the calculation results, and C(A', B') is the matrix A' and matrix B' contrast contrast function,

is the variance of the matrix A',

is the variance of matrix B', S(A', B') is the structure comparison function of matrix A' and matrix B', σ _A'B' is the covariance of matrix A' and matrix B', α, β, γ Both are to adjust the parameters among the three modules, A' is the real-time acquisition image matrix after reorganization, and B' is the standard sampling image matrix after reorganization.

In this embodiment, the specific pixel difference in the image can be calculated through the evaluation function, so as to find defects. Obviously, this formula is judged from the pixel level, so it has a strong sensitivity, so that false detection may occur. There is no missing detection phenomenon, and there is false detection phenomenon, which is caused by the incompleteness of the current image information. This is a defect of using the structural similarity algorithm. This defect has been corrected in the self-correction module.

S4. Carry out the adjustment process of the number of scanning lines according to the judgment result, and use the adjusted scanning line number information as the input image line number information in step S2, and return to step S2, as shown in Figure 6, the implementation method is as follows:

S401. Determine whether there is a defect detection image input, if so, enter step S402, otherwise, end the process;

S402. Determine whether the previous scanning line information is needed, if so, set the current line number information as the required previous line number R _pre , and enter step S403, otherwise, the imaging range of the current image is the initial range of the PCB circuit board, And enter step S403;

S403, judge whether to need the scanning line information afterwards, if so, then set the current line number information as the required number of lines R _after , and use the required number of lines R _after to complete the current image information, and enter step S404, otherwise , end the process;

S404. Using the required number of rows before R _pre and the number of rows after R _after as information on the number of input image rows in step S2, and returning to step S2,

The expression of the line number information of the input image is as follows:

ω′＝ω+R _pre +R _after

Among them, ω' is the line number information of the output image, that is, the value of the next spliced line, R _pre is the required previous line, R _after is the required subsequent line, and ω is the default value of the spliced line number.

In this embodiment, the self-calibration module aims to avoid the false detection phenomenon caused by insufficient image information by adjusting the line number information in the current image, and output the adjusted line number information to the image reorganization and splicing module to complete the self-calibration In the process, it should be noted that the information of the previous line and the next line can be adjusted according to requirements.

Through the above design, the present invention can improve the efficiency of real-time collection and detection, has high accuracy and real-time performance, and realizes the further development of PCB circuit board online detection.

Claims (9)

1. A PCB circuit board defect detection system based on a CCD camera is characterized by comprising a CCD image acquisition module, an image recombination splicing module, a defect detection module and a self-correction module, wherein the image recombination splicing module, the defect detection module and the self-correction module are respectively arranged on the same upper computer with the CCD image acquisition module;

the CCD image acquisition module is used for scanning and acquiring images in real time and transmitting the images acquired by real-time scanning and the acquired standard sampling images into the image recombination and splicing module; the CCD image acquisition module comprises a CCD camera and a CCD image acquisition interface;

the image recombination and splicing module is used for receiving image line number information transmitted by the self-correction module, respectively performing image splicing and recombination on an input real-time acquisition image and a standard sampling image according to lines according to the image line number information, and transmitting the spliced image to the defect detection module, and specifically comprises the following steps:

inputting a real-time acquisition image;

inputting image line number information: setting a default value of the number of splicing lines, and calculating to obtain a next splicing line numerical value according to the default value of the number of lines;

image splicing and recombining: according to the default value of the splicing line number, carrying out image splicing recombination processing on the standard sampling image and the real-time collected image according to lines;

the defect detection module is used for calculating the brightness, the contrast and the structure of the corresponding image according to the spliced acquired image and the standard sampling image by using a structure similarity algorithm, obtaining the difference between the acquired image and the standard sampling image according to the calculation result, obtaining the defect position according to the difference and transmitting the defect position to the self-correction module;

and the self-correction module is used for adjusting the scanning line number according to the obtained defect position and inputting the adjusted line number to the image splicing module.

2. A PCB defect detection method based on a CCD camera is characterized by comprising the following steps:

s1, acquiring a standard sampling image, setting the acquisition line number of a CCD camera and the basic setting of the CCD camera, and scanning and acquiring the image in real time by using the CCD camera;

s2, inputting image line number information, and respectively carrying out image splicing recombination processing on the standard sampling image and the real-time acquisition image according to the image line number information, wherein the image splicing recombination processing specifically comprises the following steps:

the step S2 includes the steps of:

s201, inputting a real-time acquisition image;

s202, inputting image line number information: setting a default value of the number of splicing lines, and calculating to obtain a next splicing line numerical value according to the default value of the number of lines;

s203, image splicing and recombining: according to the default value of the splicing line number, carrying out image splicing recombination processing on the standard sampling image and the real-time acquisition image according to the lines;

s3, respectively calculating the brightness, the contrast and the structure of the spliced acquired image and the standard sampled image by using a structural similarity algorithm, judging whether defects exist according to the calculation result, if so, entering the step S4, otherwise, finishing the detection, and thus finishing the defect detection of the PCB;

and S4, adjusting the scanning line number according to the judgment result, taking the adjusted scanning line number information as the input image line number information in the step S2, and returning to the step S2.

3. The PCB defect detection method based on the CCD camera of claim 2, wherein the matrix expression of the real-time collected image in the step S1 is as follows:

the matrix expression of the standard sampling image is as follows:

the method comprises the following steps of A, B, epsilon, y and eta, wherein A is a matrix of a real-time acquired image, B is a matrix of a standard sampling image, epsilon is an initial scanning position, x is the number of lines of each scanning, y is the number of columns of each scanning, and eta is an initial value of standard sampling image interception.

4. The CCD camera-based PCB defect detection method of claim 2, wherein the expression of the next stitch line value in step S202 is as follows:

ω′＝ω+R _pre +R _after

where ω' is the next splice line value, R _pre For the desired forward movement, R _after ω is the default number of stitching rows for the desired next row.

5. The PCB defect detection method based on CCD camera of claim 2, wherein the matrix expression of the real-time collected image after the recombination in the step S203 is as follows:

the matrix expression of the recombined standard sampling image is as follows:

wherein, A 'is a recombined real-time collected image matrix, B' is a recombined standard sampling image matrix, and R _pre For the desired forward movement, R _after And omega is a default value of the number of splicing lines, x is the number of lines in each scanning, y is the number of columns in each scanning, and eta is an initial value of standard sampling image interception.

6. The CCD camera based PCB circuit board defect detecting method of claim 2, wherein the step S3 comprises the steps of:

s301, respectively calculating the brightness, the contrast and the structure of the spliced real-time collected image and the standard sampled image by using a structural similarity algorithm;

s302, fusing the brightness, the contrast and the structure of the real-time collected image and the standard sampled image in proportion to obtain an evaluation function;

s303, judging whether the evaluation function is larger than a preset detection threshold value T or not _d If so, finishing the detection so as to finish the defect detection of the PCB, otherwise, marking the defect position in the current acquired image, outputting the current defect detection image, and entering the step S4.

7. The CCD camera-based PCB circuit board defect detection method of claim 6, wherein the expression of the evaluation function is as follows:

F(A',B')＝[L(A',B')] ^α [C(A',B')] ^β [S(A',B')] ^γ

wherein F (A ', B') is an evaluation function,. Mu. _A' Is the pixel mean gray value, μ, of the matrix A _B' Is the average gray value of pixels in the matrix B', N is the total number of pixels, x _i Is the value, y, of the pixel corresponding to the matrix A _i Is the value of the pixel point corresponding to the matrix B ', i is the subscript of the corresponding point in the matrix A', sigma _A' Is the standard deviation, σ, of the matrix A _B' Is the standard deviation of the matrix B',l (A ', B') is a luminance contrast function of the matrix A 'and the matrix B',

which is the square of the average gray value of the pixels of the matrix a',

is the square of the mean gray value of the pixels of the matrix B', C ₁ ，C ₂ ，C ₃ All are stability parameters for increasing the calculation result, C (A ', B') is a contrast ratio function of the matrix A 'and the matrix B',

is the variance of the matrix a' and,

is the variance of matrix B ', S (A ', B ') is the structural contrast function of matrix A ' and matrix B ', σ _A'B' The covariance matrix A 'and the covariance matrix B' are all parameters for adjusting the three modules, A 'is a recombined real-time collected image matrix, and B' is a recombined standard sampling image matrix.

8. The CCD camera-based PCB defect detection method of claim 2, wherein the step S4 comprises the following steps:

s401, judging whether a defect detection image is input or not according to a judgment result, if so, entering a step S402, otherwise, ending the process;

s402, judging whether the previous scanning line information is needed or not, if so, setting the current line number information as the needed previous line number R _pre And step S403 is carried out, otherwise, the imaging range of the current image is the initial range of the PCB, and step S403 is carried out;

s403, judging whether the subsequent scanning line information is needed or not, if so, setting the current line number information as the subsequent line number R needed _after And using the desired number of columns R _after Completing a current imageInformation, and step S404 is entered, otherwise, the flow is ended;

s404, the required previous row number R _pre And the required number of post rows R _after As input image line number information in step S2, and returns to step S2.

9. The CCD camera-based PCB circuit board defect detecting method of claim 8, wherein the expression of the row number information of the input image in the step S404 is as follows:

ω′＝ω+R _pre +R _after

where ω' is the number of lines information of the output image, i.e. the next stitching line value, R _pre For the desired forward movement, R _after ω is the default number of stitching rows for the desired next row.

Images